Watershed method for controlling cashflow mapping in value at risk determination

• US 6,122,623 A
• Filed: 07/02/1998
• Issued: 09/19/2000
• Est. Priority Date: 07/02/1998
• Status: Expired due to Term

• Current Assignee: Financial Engineering Associates, Inc. (ION Investment Group Ltd.)
• Original Assignee: Financial Engineering Associates, Inc. (ION Investment Group Ltd.)
• Inventors: Garman, Mark B.
• Primary Examiner(s): Stamber, Eric W.
• Assistant Examiner(s): Jeanty, Romain
• Application Number:
  US09/109,110
• Time in Patent Office: 810 Days
• Field of Search: 705/35, 705/36, 705/37, 705/38
• US Class Current: 705/36.00R
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

A system, computer implemented method, and software product provide for the correct allocation of cashflows to enable accurate determination of value at risk with respect to income and balance sheet risk for transactions portfolio including transactions occurring in different fiscal periods. The computer implemented method includes establishing watershed variables, such as watershed dates, and partitioning both cashflows derived from the transaction portfolio and the vertex set of market risk data into distinct subsets. The partitioned cashflows are allocated, using a regular allocation function, onto individual ones of the partitioned vertex sets. The partitioning and allocation correctly segregate cashflows with respect to the fiscal periods to which they contribute to the value at risk. The allocated cashflows are then each separately treated by a value at risk computation. A system includes a computer, database of transactions, networked or local access to market risk data, and a software product executing the computer implemented method. The software product may include a module for shredding transactions into cashflows, a module for partitioning the cashflows and vertex sets, a module for performing the regular allocation of partitioned cashflows, a module for performing the value at risk computations.

Claims

1. A computer implemented method of determining a value-at-risk measure for each of a selected set of transactions, the computer implemented method comprising:

converting the transactions into a set of cashflows;

storing a set of vertices, each vertex including an edge value for each of a plurality of edges, each edge having a unit type, wherein selected ones of the vertices form a canonical vertex set;

receiving a set of N watershed variable values for at least one of the plurality of edges;

converting the N watershed variable values to a same unit type as the edge;

partitioning the values of the edge into N+1 partitioned edge value sets according to the N watershed variable values;

partitioning the vertices of the canonical vertex set according to the N+1 partitioned edge value sets to produce N+1 partitioned vertex sets, each partitioned vertex set including at least one vertex;

partitioning the set of cashflows according to the N+1 partitioned edge value sets, to produce N+1 partitioned cashflow sets, each of the N+1 partitioned cashflow sets including at least one cashflow;

for each j^th (j=1 . . . N+1) partitioned cashflow set, allocating the cashflows within the j^th partitioned cashflow set onto only the vertices in a corresponding j^th partitioned vertex set; and

for each partitioned vertex set, determining a value at risk for the cashflows allocated unto the vertices in the partitioned vertex set.

2. A computer implemented method of determining a value-at-risk measure for each of a selected set of transactions, the computer implemented method comprising:

converting the stored transactions into a set of cashflows, each cashflow associated with a date, wherein there is an ordered set of dates associated with the cashflows;

partitioning the ordered set of dates into N+1 partitioned date sets according to a set of N dates;

partitioning a set of vertices according to the N+1 partitioned date sets and a date edge of each vertex to produce N+1 partitioned vertex sets, each partitioned vertex set including at least one vertex;

partitioning the set of cashflows according to the date of each cashflow and dates in each of the N+1 partitioned date sets, to produce N+1 partitioned cashflow sets, each of the N+1 partitioned cashflow sets including at least one cashflow;

for each j^th (j=1 . . . N+1) partitioned cashflow set, allocating the cashflows within the j^th partitioned cashflow set onto only the vertices in a corresponding j^th partitioned vertex set; and

for each partitioned vertex set, determining the value at risk measure for the cashflows allocated onto the vertices in the partitioned vertex set.

3. A computer implemented method of determining balance sheet value at risk and income statement value at risk for a portfolio of transactions, comprising:

converting the transactions into a plurality of cashflows;

partitioning the cashflows into a first cashflow set and a second cashflow set with respect to a selected date;

partitioning a set of market risk data into a first market risk data set and a second market risk data with respect to the selected date;

mapping the first and second cashflow sets onto the respective first and second market risk data set to produce first and second mapped sets of cashflows;

determining the income statement value at risk measure from the first mapped set of cashflows; and

determining the balance sheet value at risk measure from the second mapped set of cashflows.

4. A computer implemented method of determining an analytic variance-covariance value-at-risk measure of balance sheet and income statement risk for a portfolio of transactions, comprising:

converting the transactions into a plurality of cashflows;

partitioning the cashflows into a first set of cashflows occurring on or before a selected date, and a second set of cashflows occurring after the selected date;

partitioning a variance-covariance matrix of market risk factors associated with a set of predetermined constant maturity tenors (CMTs), into a first submatrix of market risk factors having CMTs occurring on or before the selected date, and a second submatrix of market risk factors having CMTs occurring after the selected date;

mapping the first set of cashflows onto the first submatrix of market risk factors;

mapping the second set of cashflows onto the second submatrix of market risk factors;

determining a first value at risk measure of income statement risk from the first mapped set of cashflows; and

determining a second value at risk measure of balance sheet risk from the second mapped set of cashflows.

5. A computer implemented method of determining an analytic variance-covariance value-at-risk measure of balance sheet and income statement risk for a portfolio of transactions, comprising:

converting the transactions into a plurality of cashflows;

partitioning the cashflows into a first set of cashflows occurring on or before a selected date, and a second set of cashflows occurring after the selected date;

partitioning a RiskMetrics data set of vertices and a covariance matrix into a first subset having vertices occurring on or before the selected date, and a second subset having vertices occurring after the selected date;

mapping the first set of cashflows onto the first subset of vertices;

mapping the second set of cashflows onto the second subset of vertices;

determining a first VaR measure of income statement risk from the first mapped set of cashflows; and

determining a second VaR measure of balance sheet risk from the second mapped set of cashflows.

6. A computer implemented method of determining an analytic variance-covariance value-at-risk measure for a portfolio of transactions, comprising:

converting the transactions into a plurality of cashflows;

partitioning the cashflows into a first set of cashflows occurring on or before a selected date, and a second set of cashflows occurring after the selected date;

providing a set of predetermined constant maturity tenors (CMT), each CMT associated with a price volatility and a covariance with the other CMTs;

partitioning the set of CMTs into a first subset having CMTs maturing on or before the selected date, and a second subset having CMTs maturing after the selected date;

mapping the first set of cashflows onto the first subset of CMTs;

mapping the second set of cashflows onto the second subset of CMTs;

determining a first VaR measure of income statement risk from the first mapped set of cashflows; and

determining a second VaR measure of balance sheet risk from the second mapped set of cashflows.

7. A computer implemented method of determining an analytic variance-covariance value-at-risk measure for income statement risk for transactions occurring prior to a selected date, and a value at risk measure for balance sheet risk for transactions occurring after the selected date, the method comprising:

providing a set of predetermined constant maturity tenors (CMT), each CMT associated with a price volatility and a covariance with the other CMTs;

converting the transactions into a plurality of cashflows, each cashflow occurring on a date;

creating a first set of allocated cashflows by;

determining a last CMT occurring prior to the selected date;

allocating all cashflows occurring on the selected date, or between the selected date and the last CMT only to the last CMT; and

regularly allocating all other cashflows occurring prior to the last CMT to the other CMTs prior to the selected date;

creating a second set of allocated cashflows by;

determining a first CMT occurring after the selected date;

allocating all cashflows occurring after the selected date and on or before the first CMT only to the first CMT; and

regularly allocating all other cashflows occurring after the first CMT to the other CMTs after the selected date;

determining a first value at risk measure of income statement risk using the first set of allocated cashflows; and

determining a second value at risk measure of balance sheet risk using the second set of allocated cashflows.

8. A computer system for determining an analytic variance-covariance value at risk measure for a selected set of transactions of a portfolio of transactions stored in a database, the system comprising:

a shred module that generates a set of cashflows for each of the selected transactions;

a partition module that receives at least one user defined watershed variable, and converts the watershed variable to a type of value that is comparable to a corresponding type of value of the cashflows, and that partitions a set of risk characteristic vertices and the cashflows using the watershed variables into respective vertex sets and cashflow sets; and

a value at risk module coupled to the partition module and the database, that receives the partitioned vertex sets and the partitioned cashflows sets, and covariance data corresponding to the risk characteristic vertices, and computes a value at risk measure for each partitioned cashflow set.

9. A computer system for determining an analytic variance-covariance value at risk measure for a selected set of transactions of a portfolio of transactions stored in database, the system comprising:

means for converting the stored transactions into a set of cashflows, each cashflow associated with a date, wherein there is an ordered set of dates associated with the cashflows;

means for partitioning the ordered set of dates into N+1 partitioned date sets according to a set of N dates;

means for partitioning a set of vertices according to the N+1 partitioned date sets and a date edge of each vertex to produce N+1 partitioned vertex sets, each partitioned vertex set including at least one vertex;

means for partitioning the set of cashflows according to the date of each cashflow and dates in each of the N+1 partitioned date sets, to produce N+1 partitioned cashflow sets, each of the N+1 partitioned cashflow sets including at least one cashflow;

means for allocating the cashflows within each j^th (j=1 . . . N+1) partitioned cashflow set onto only vertices in a corresponding j^th partitioned vertex set; and

means for determining for each partitioned vertex set, the value at risk measure for the cashflows allocated onto the vertices in the partitioned vertex set.